Training

[Guyuena]

I train the model.py in my machines but the pacc1 is less 0.05 all the time. I don't know that Is right or wrong?

[lightvector]

What data are you training it on? If you're starting from very early in training from scratch, then it will of course be hard to achieve any high prediction accuracy because random play is impossible to predict accurately. If you're training on data from a mature net or other games, then that does seem a little low, although 6100 steps is not much training either.

[lightvector]

Looks like in #639 - you trained much longer and still it doesn't predict much better. What data are you training it from? You can also try asking for help in https://discord.gg/45EWcZu7

[Guyuena]

ok,thank you